Getting Started With OpenStack

shinsen I think your honor I'll be on good hey good morning everyone thanks for coming after the party you guys are hearty souls I like that alright um we're going to we're going to start going to give a little intro to intro of who we are and then dance and the gift some kind of up updated or some instructions that you guys in case you guys haven't seen it on how to kind of catch up laptop prepped with a vagrant imaged so you can kind of follow along during the workshop portion so well mater is going to do the intro and I'm going to give a quick overview kind of a quick snapshot of what is OpenStack and where it stands today and then Dan will take you through kind of the internals and how everything works so before I get started though let me ask how how many you have new to OpenStack you know what you have it okay how would you classify good okay then this is the right workshop for you great alright let's get started so ah let me introduce myself my name is Ken hoi I'm currently at Rackspace as a as they OpenStack evangelist and I've been involved with OpenStack for about three years and I'm Dan radies I'm on the OpenStack team at Red Hat been a Red Hat for about eight years and working on OpenStack for three and a half now I think I counted seven summits that I've been able to attended me a part of the community the so Ken was talking about the vagrant file if you didn't see in the abstract of this session there's a link out to a quick web page that we put together that references a vagrant file so while he's given a little bit of the introduction of OpenStack and how it came to be in some of its history pull down that vagrant file and go ahead and see if you can get it running to get that virtual machine running because the the vagrant file that's out there is the same exact one that I'm going to present off of so you can follow along with it it takes a little while to go to finish doing the installation so if you don't quite get it done here we'll also have a link to the slides at the the end of the session as well so you can take the vagrant file and the slides later and then work through it again so everything here you'll be able to do now as we're working through it as well take it with you and give it another try you know on your own time as well okay great so Dan just gave you permission not to listen to me while you're busy getting your favorite Vox up that's okay so so yeah go so go on to the schedule and go on the apps track and somewhere in the bottom I think there's a link to the figurin box that Dan put together so okay while you're doing that let's I'm going to give a great quick kind of history of where OpenStack came from so this email you're looking at is the actual email that was sent by an executive at Rackspace at the time in 2010 to NASA seats the CTL NASA at the time is venture essentially inviting them to come alongside Rackspace to create a new open source cloud platform so so history of that is that Rackspace up to 2010 have been running their own public cloud but but it was not it was like proprietary piece of code and it had scalability challenges so when they got to a certain point where they knew they couldn't scale anymore they decide they want to rewrite the entire thing another point they made a decision to use Python as the underlying programming language and at this at the same time or roughly the same time I was happening NASA had an initiative to create a private cloud and they decided they wanted to see they could pick something that was open-source if possible and they evaluate some of the current the platforms are out there at that time like I think cloud stack eucalyptus and others and decide what didn't quite fit their needs so they decide they would build their own public cloud as well I mean on private cloud and they in the pendant lis of Rackspace also chose to use Python as their program language so the exact leaf that who sent the email you saw in the previous slide read about it contacted Chris Kemp who is the CTL NASA at the time and basically said hey do you want to work together and so all that is where OpenStack came from and and one of the one students they may was even though they had these two teams working together they still were limited somewhat limiting resources like NASA is not it's not a software company per se and while Rackspace had allows software developers it is by nature a match was at that time and managed a hosting company so decide the best way to grow OpenStack quickly was to open-source it and give it to the community as community people who could then write code and kind of build up the project so this is I'm not sure let's try a little outdated but it kind of shows you in the six years time how quickly the open-source project has grown from really just so those two entries to the point where you have hundreds of companies spanning multiple geographies and you also you can also see that OpenStack is now starting to grow in his adoption so this is a list of companies that that have we they may have adopted OpenStack much at the very beginning but these are the ones have come out in the last year and publicly said hey we are in fact using OpenStack and you can see it's it's across multiple it's not just a science research it's not just public cloud companies it's not a true enterprise adoption so since you guys are new it will give a very quick definition of what is of oviya what is OpenStack so this picture you may have seen from the foundation OpenStack Foundation website so basically if you think about what what data center infrastructure was like before cloud computing it's basically a lot of silos of physical hardware like physical servers networks and storage and then somewhere along the way we started figuring out how to virtualize much of that and started being it Claypool's of resources the problem though was each of those silos had to be managed in provision separately and it requires someone who had a lot of knowledge about how to set up a server set up an you know piece of networking gear storage so the idea behind cloud computing which is where OpenStack fits into is what if we could what if we could figure out a way to kind of automate all that management and provisioning of those virtual resources and make it really easy for a end user who doesn't necessarily know much about storage or about networking to be able to to provision his their own resources if I having to go through an IT person right the idea to be able to do things very quickly so if you think if you look here that basically are these southbound api's that talk to the infrastructure to manage it and that's basically any virtualization platform should be able to do what makes OpenStack a cloud a true cloud computing platform is actually the northbound API right the fact that again a developer through ap is or through a web portal could actually say hey I want ten machines that now network together than this way with this amount of storage and never have to talk to an operator never have to talk to a storage admin and say hey I need you know ten two terabytes of storage tomorrow and the idea also if you look at this picture one of the difference between OpenStack and some other platforms is that it is a very loosely couple architecture in other words all the components within OpenStack from from the user experience the dashboard to the storage to the networking the servers the it isn't one monolithic system that makes system calls to each other right everything is done through API so Evie Evie Evie components really zone project it's kind of own own program and we're just relying on these were open REST API to talk to talk to each other and the idea is this way you can change out different parts you can you can scale out different components without necessarily negatively impacting other components that's in all these components that he has on this slide is what we're going to go through and the hands-on part so we're going to identify other different pieces on how they interact with one another yeah so this is just kind of giving you a kind of overview of how the different components tie together dan will actually walk through what each of these components do within an OpenStack cloud platform so at the end of the day the the go of OpenStack right is to sit be able to allow developers to create applications that that they can build on a platform that they can take in self-service right create and then rapidly scale write that is fundamentally what OpenStack is designed to do let developers move really fast and grow applications to scale this is a example of a reference architecture of what a tip if a smallish or typical OpenStack implementation might look like this is actually based on real head OpenStack platforms reference architecture and you see here again that's different the OpenStack counts spread out across many many components so many servers in order to deliver this cloud computing platform talk quickly about some consumption models different ways that you can actually consume or use OpenStack today so they have three primary ways one is a public cloud private cloud and then what we call the middle ground which is private cloud as a service so public cloud pretty easy to understand it's what Rackspace has is what dreamhost has its what Amazon use it has even though they are not running an open stack so let's share the infrastructure that anyone can get access to private cloud distro again it's a it's that the design is to have a share a private infrastructure infrastructure that no one else has access to besides your company and there's a number players that do that and then private cloud is a service it's kind of a middle kind of a middle ground where you can do some trade-offs raise it still it is single tenant but you don't manage it right because one day when the value props of a public cloud is you don't have to run this cloud platform as an operator you just consume it as a developer someone else handles it and the downsize that's all you control someone else has to run it for you any and you're sharing it with other people a private cloud you get you have exclusive access you get to control it the problem is you have to you have to operate on a day to day basis and then the private clouds of service is that middle ground that says consume it like it's a public cloud but you're basically handing off management to someone else like it's a bubble even though it's again running inside maybe running inside young data center so a couple of the big players this is just a sample then many if you look on the OpenStack org marketplace web page there is many many more vendors that involve I'm just kind of laying out some of the big key ones I know of that then you'll notice then some you'll see that in some cases as you as you look on that website that some vendors actually play across two or three of these spaces so this is the slightly marketing piece you know how that will make it very long so as I mentioned I'm from Rackspace so Rackspace is a couple of different approaches to to presenting OpenStack so I talked about the public cloud because that's basically what a public cloud is built on today is built on OpenStack from a private cloud perspective we we actually don't have a distribution per se we have we offer that private clouds of service so everything we do is as I serve as a service offering that we manage whereas our public cloud or is a single tenant private cloud for a customer and we offer that in two flavors right now one is a the private clouds using just using the upstream code rate that's running on a Bunter or we we can also do private clouds of service using Red Hat OpenStack so customers have either option and then Red Hat up to a few months ago really had the one option of his a distribution that you can download and read a book can help you set up and deploy but primarily you run it yourself right so what's new now is because of what we're doing around wrapping our managed services around Red Hat OpenStack essentially red now has the cat has both the distribution offering and also this private clouds of service offering okay that was the marketing piece let me talk a little bit about it some ways you can learn OpenStack obviously you guys being this workshop it's a great way to get started but they love there are other things you can do once you get back to whether you're you know your home is for you and be able to continue learning OpenStack so a couple of things one is there's a bunch of resources you can look at OpenStack foundation is probably the best place to start it has documentation have videos so you can kind of learn more about what OpenStack is and how to use it few books that are out there's an open stack pile computer cookbook that was win by a couple of Rackspace engineers that it's a great way to use I think they also use vagrant we will spin up a kind of do what we're doing here spin up a one no a multi no environment and be able to play with it OpenStack essentials is actually the book that Dan wrote that does a similar thing so if you like the workshop today go by his book it's it's my suggestion Kayla and I have done this workshop what four or five five it's not fifth time a together so we've kind of become partners in crime yeah this OpenStack essentials book is essentially a print form of the hands-on part of this presentation that I'll do so a lot of the same exact stuff that I do right here today this is a print form that will walk you through and give you more in-depth information about the different pieces it's actually the first draft of the second edition has just been completed last week so hopefully in the next week or two here we'll have the second edition published so you're welcome to get the first edition it has lots of great information there's a few extra updates and additional information that's been added so if you want to wait a couple weeks for the second edition okay that's good to know yeah all right there's other books that have been out for a long time there was an F there were no resources for learning OpenStack then the web site it was a lot of googling or googling around and finding things that actually didn't work because it was outdated so since open stacks got more mainstream there's many more kind of real books that have been written so I forgot to put this in the deck but it is we all racks is unfortunate because of our involvement with OpenStack for since this very beginning we've had a lot of experience and a lot of people have worked on it and a lot of have gone on the write books so that right now there are three or four books that are including the one in the screen that's offered by Rackspace employee-employer engineers and we're actually giving them all away at I don't know how many of you seen the Rackspace Cantina kind of the restaurant okay so this afternoon I think starting I want to say starting at three yeah five three we're going to be giving away three of the books that have been written by Rackspace engineers so you just have to go to the cantina get online they'll hand out the book if you want they can sign it and you can actually talk to them and ask questions ask the office questions so a couple several of those people are actually req other actually current Rackspace engineers so one of being like a for example a network probably the best RAC OpenStack networking book that's out there today it has been by one of our engineers and he's going to be signing the book and giving them away to people at the cantina later this afternoon sir alright lastly I think it's the last slide for me if there's likely a user group in your kind of where you live or very close hopefully there is if there is that these group tend to meet every month or every other month and they go over you know thing other technologies involve of OpenStack o2 sometimes they do hackathons or just workshop so I encourage you to go to the OpenStack tower community website find a user group near you and get involved with that community on a regular basis the other thing is there are the foundation started doing something called OpenStack days which is basically think of it as like a mini summit that's done regionally so it's historically that's been done internationally last couple years they did a vertical OpenStack in Silicon Valley and then later this year that we will have the first OpenStack day East which will be in New York City so again tend to be one or two day events where would you have a keynote and you have breakouts just like you would at a full-blown OpenStack summit ok so that let's get going I'm going to let Dan take over from here you need to click it let me borrow that data start out with and I'll stay here just if you guys have questions along the way between then and I wish your bid or I answer them okay so it had the way that redhead operates as a companies but every product that we have also has a associated community project so everything is open-source that we have everything that we write goes first into an upstream if possible and if it doesn't go directly into the upstream will carry the patch internally until we can get it into the upstream so as Ken mentioned before there's Red Hat OpenStack platform which is our supported enterprise product but on the community side we have our do so our do is our community supported distribution of OpenStack and that's what we're going to use here today that's what if you've if you've been able to get your virtual machine up and running with vagrant it has gone out to RDO and installed our do into this virtual machine and that's what we're going to walk through our do we take the upstream source and we package it into RPM and then we give it to you so what you get in our do is directly what comes from the upstream if you're installing Red Hat Enterprise so our Red Hat they just changed the name so I'm trying to learn how to say it correctly Red Hat OpenStack platform if you're using that it's it is supported by the company and so we oftentimes will carry patches for customers so there's a small Delta between what's in the community distribution are do versus what's in the enterprise distribution OSP OpenStack platform and so but if you look at the two really the only difference between the two of them on a in a big picture is is just the branding of it that when you install OSP it's Red Hat branded and when you install our do there's no branding it's just vanilla OpenStack so that gives you a little idea of what we're installing and using today are do project org down there at the bottom is where all kinds of documentation where you go to get started if you wanted to not use the vagrant file that I've given you if you want to do a larger installation or learn how to do more than just a little sample installation lots of documentation and materials there so this is a picture of the components that we're going to go through I'm going to quickly touch on each one of them and then we'll get started actually looking at each of them Keystone is identity so this is where you're going to be able to do authentication and all the services are registered there so everyone has an identity and needs to be authenticated with one another not just the end-users but all these components together after we do identity we'll look at glance glances image management so when a VM starts it has to have a disk behind it and so what we do is we pre build the images and load them into Glantz so that when the VM starts it goes out to glance and pulls a copy of the image that's been pre-built and then instead of you having to go through the whole installation process it's pre installed and ready to go and Nova boots it up customizes the networking and identity inside of it and then you're ready to run so it's a way to quickly get those VMs up and running once you before you can get that instance actually running you need the disk image from glance and you also need a network to attach it to so that next we'll go to Neutron which is OpenStack networking and we'll look at creating a virtual network to attach it to then once you have that image available and the network available and the identity to get you into OpenStack then Nova is our compute component and Nova is going to take all those pieces and put them together talk to the hypervisor and actually launch that VM using the image using the network that has been provided to it once the instance is up and running the instance in in cloud computing or in OpenStack the paradigm is elastic computing and the idea is that these VMs are intended to be kind of disposable that if you have multiple VMs that are working together to run an application and one goes down instead of working really hard to try and keep that one backup and figure out what went wrong with it while end-users are waiting for that capacity to come back instead you just kind of slice it off and spin up a new one because it's so quick and inexpensive to do that and so doing that the disk underneath it is ephemeral it doesn't last it gets thrown away with the image if it gets axed so cinder is our volume service we can attach a persistent block device out to those VMs to write information to that block device that way if the instance gets axé you're able to reattach that block volume device to another instance so that you can continue using the information that you've been using this isn't shared storage it's block storage so it's a one-to-one relationship between the the drive and the the volume and the instance there is a shared storage called Manila which we won't get into so if you need shared storage you can do that as well object storage is very simple storage so it's it's simple content object so instead of working at the block level as in presenting a volume as a disk to the VM instead you use an API and you pass content with kind of a name value sort of mentality that you can this is the name of the object that I want to store and this is the content that's going to go in it or this is the name of the object that I want to pull and this and it gives you the content in it so it's it's very basic file storage but it can be very powerful there's there's websites that use object storage to run their entire system because of how flexible and how simple it is as well as as their software-defined and behind it so you can use commodity servers to do replication and mirroring and there's a lot of power behind object storage it's not just simple file transfer back and forth yeah and you've probably um if you've been at this how many give this is this your how many people here this for them this is their first summit Wow okay allow you all right so you might have heard that there's a lot of these all these other project names I haven't mentioned like Magnum in projects like that so OpenStack actually is made up of I think currently over 50 projects which can be dizzying obviously we're not showing all 50 projects these are what I would consider to be more core projects in other words what are the kind of what the minimum things I need to get have in order to say spin up a OpenStack cloud so this is how so what Dan and I are going through the kind of core stuff and then all these other forty some other odd other projects are useful projects then what things that you layer on to your core your base OpenStack cloud we just don't have time to touch all those in the hour and a half they give us this morning and then all fit in one fly other probably also be bored to tears by the 11th okay so the last one at the top here is the dashboard it's based on a project called horizon which is more of a framework so dashboard and horizon are are somewhat analogous in the component that you're working with the the technical difference is that horizon is the framework underneath it that the dashboard the web interface is built on top of so that's the first thing we're going to do here is is connect to the dashboard the web interface and that's where we're going to be today working through these concepts so that you can learn and understand how they work together through this this graph we're going to go in and create each of these virtual resources and attach them together to get a virtual machine up and running an OpenStack everything in OpenStack is built module or modular modular modular and the dashboard is no different so as Ken mentioned there's almost 50 project at this point and so as each of those projects come through the dashboard is committed to working hard to get each of them to have web support dashboard support in it and so it has to be built module II that a new project comes online our project that previously hasn't had a web interface comes into the dashboard and says we're ready to have our our web interface there or we've done this work for it they've created another module that they can drop in and integrate in and so it it makes it quick in how it ends up being able to be more projects can be added to the dashboard so in that in our abstract there was the web link that had a link out to the vagrant file if you're able to get that up and running and installed and this is kind of where we start getting into that so vagrant up is going to do that installation if you haven't done it yet it's going to take a little while for the OpenStack to actually be installed if you've already done it and then did a vagrant halt like I'd put in the instructions then you should be able to vagrant up again and it'll come right back to where it was before when when you shut it down previously the next thing we need to do then is connect to the dashboard and try and log in using it so vagrant SSH will log you into the command line of that VM that you've brought up sudo - I will change you to the root user and we're going to talk a little bit more about installation methods at the end the installation method we use is called pack stack here and it's good for kind of one-off simple demo like environments which is why it's being used here and when it installs the Keystone RC admin file is dropped into the root users home directory so if you cat that file out and list out the contents in it the administrator username and password that get generated for you are put in there and then finally well this is the web URL that vagrant has helped us to present so you should be able from your web browser once OpenStack is completed installing connect to this 192 168 37 2 and the dashboard will get appended if you just connect to the IP address it'll redirect out to the dashboard let's do a quick check how many of you here have gotten to the point where you've been able to figure it up okay how many you're still working on getting the vagrant but okay great okay and so as I mentioned before - if you're having trouble getting it to work or it's still running catch up as you can but also I'll have the slides for you later and you know take the file with you so you should be able to use this after the presentation as well and we'll give you Ken's email so you can send all the sessions to him if it doesn't continue to work yeah it's in the abstract for the session there's there should be a link in there you like the schedule the the summit schedule yeah go go find the the the session description in the schedule and the link is down there at the bottom okay so connecting to the dashboard so if you connect to that URL that I put in there you're going to see a screen that looks like this and then if you SSH in oops let me get my vagrant running I'm sorry the font bigger oh yeah I'm not doing anything that's important right now I'm I'm restarting vagrant there we go here we go can you see that it's a good okay so there you go you see if my vagrant was just suspended so it was still running but it was it was in a suspended state so I resumed it and then vagrant SSH so now I'm on that VM that vagrant created for us and then sudo - I and cat Keystone RC so you'll see here that there's OS username admin so that's the generic admin username that's generated for us and then OS password right underneath it is a randomly generated password and so that's what we need to get into the dashboard so I'm going to copy that password and login as admin and now I'm logged in as the administrator user to the dashboard so let's keep going from there oh so Keystone Identity Management is next the idea here is that in the install that we've done we've made it a centralized identity service and a centralized catalog of services what this means is that all the users and all of the components within OpenStack can call in to Keystone and ask how to connect to the different components so in particular when if you use the command line and or the command line or the web interface every time you make a call to create a virtual resource you have to connect to one of the components that we're looking at today and so when you connect to that component you have to be authenticated and then if that component needs to talk to another component to create virtual resources or associate them in some way those components have to authenticate to each other so there's tokens and user names and passwords that are all passed around from the users and the services and so the identity service is the users and the catalog of services is that you can ask Keystone how do I connect to service Nova or to service glance and it will respond to you with a connection URL so that you can make your connection to that service yes one one way to kind of one illustration it may help in your head is think of as a think of all the components that you need to spin up some resources like that the spinning up a server spinning up a some switcher swimming up some storage think of each of those as having workers that you do that work for you and when the cert server worker needs to storage it needs to essentially present a badge that says hey this is who I am and I'm authorized to ask you to present me some storage right that's essentially what Keystone is is a waited for these workers to present badges the authorized badge each others say I need a resource from you so that I can pull it all together spin up a resource yeah definitely and then for your authentication options these are just a couple but Keystone can also have other identity management or authentication systems plugged into it so the users can be connected to LDAP or AD or username password token ooofff you know all if you're familiar with Apache remote user where Apache also supports a bunch of other authentication schemes you can plug into Apache and rely on that remote user login methodology and Keystone will also recognize that so there's you're not tied to user name password in the way that we've done in this demonstration here so now if we create a user and logged in as the administrator user so I can manage users I'm going to click on the the users link on the side here can you guys see that little bigger would begin and then well now my buttons been pulled on and then up at the top here there's a create user the top right over here there's a create user button so you click that and it's going to give you a big dialog that you can fill in all the information so I'm going to put my name in here and I could describe myself I guess in the description so what's happening here is Dan's logged in as when you when he says is in admin he's basically the global super user for the foot that in OpenStack cloud and Dan we'll talk more about later but there's a concept of in that cloud you can have multiple tenants that have resources only available that you can assign specific resources to that tenant and then within each tenant you can give you can create users I only have access to that tenant so you don't have to give everyone that using your cloud the super user rights and just to make things super confusing OpenStack has mixed the word tenant project so if you hear tenant or you hear project they're the same thing it's just on the command line it started with tenant and then in the dashboard they started using project and now they're starting to switch to using project on the command line as well so you can see here we're at primary project the idea in Keystone is that you kind of have a triangle of things that are important you have a user name you have a project that the user will live in and then you have a role that the user is associated with the project so all of your virtual resources that get created have to be created inside of one of these projects and the user is no different a user gets assigned to a prod in general if there's a group of people that are all together then you will name the project something relative to the users but if it's a project specifically for that user the standard is kind of to just create the project name with the same name as the user so I'm going to come in here and do create a project and if I had more members to add to it I could do that on this tab but I'm just going to create project so I created project with my name that matches my user name and then what the dashboard does for you is brings you back to the same screens that you're at so all the information I'd already filled out is there and now my primary project has already filled in there when I click this plus button here and went to that other dialog I switched out of the create user and I actually created a new project object and then after that project was created I came back into the create user and so now it's associating it and then at the bottom the role mem is as member here and member is a generic non administrative role for you to be in your tenant yeah so is it in a production environment just you would likely set up you know let's say let's say developers are going to be using this you may create a tenant for every for every software project and then a user could be a member of one or more of those projects software product development projects and so in each project you can have in one place you could have an admin wrote and in another project that same user could have just a member row so the pen so there's a lot of flexibility in what you can do and then later we're going to look at using swift and Swift has a special role that it needs to be connected to so I'm going to come in here you can see the project that I created is down at the bottom that matches my username and then I'm going to go to manage members and you'll see that my name is in here and I'm a member role but I can also add myself as a swift operator and so for us to be able to do Swift later we have to add this role to my user within the project there is some configuration that you're able to automatically make all users in project Swift operators it's not configured by default in pack stack what we're using so I'll save that and now I'm a member and a swift operator so we can later use Swift so at this point I have a user I have a project and I have a role in my project so I can log out as the administrative user and then log in as myself my non privileged user and the first thing to notice here is a non privileged user is that there's no administrative panels here so as you're interacting with this log back in as the administrator log back in as yourself and note the difference of here I've got project and compute and network object stores so my basic virtual resources that I'm managing and then in the administrative there's also user management and an administrative panel that lets you globally manage all of the resources within the within this cluster so now that we have a user let's start kind of working towards getting a virtual machine up and running the first thing we talked about was the image management that glance houses these pre-built images the idea again being that we don't want to have to sit and wait for the installation to run for each of these virtual machines when they want launch that we can it's very boilerplate to do this installation so if we do the installation ahead of time and put a generic image into glance then when we launch that image all we have to do is do a couple little tweaks to make it unique in its networking information or its identity on the on as a machine and then we can it's very quick to pull that image and boot it and start running from it so glances image management it's a registry for these disk images so we import the disk images into glance and then you're able to recall them and share them across your cloud there's lots of these images pre-built for you on the internet so if there's a particular distro or flavor of OS that you want to use you know go search on the web for a cloud image or OpenStack image for the particular district one of these images pre-built for you that you can download and put it in so looking at adding an image instead of trying to distribute this image to everybody and get it so you can download it I went ahead and had it imported for you so if you go to images in this screen you can see that there's a serious image serious is an operating system that's built for developmental testing purposes so it's very insecure as an as a operating system and it's not recommended at all for use for anything but basic testing and demoing and the reason it's great for demoing and testing is that it's only 12 Meg's so if you see the size over there on the right is it's a teeny tiny little image and that's because there's not much in it but for instance if you wanted to say put Fedora or CentOS or something we could say cloud fedora and it comes right up with download class of a word yeah somewhere so here we could if I download this Fedora image here that that's an image that we could pull directly into OpenStack you might have one already downloaded so in my open stack here we say create image fedora you can actually give it a URL so this image location that's here if I had just put the direct URL to that image it would let me pass that in and then it would download it and it would import it I'm going to see if I have a file waiting for me yeah there you go so I've got a fedora image cloud image here so I've selected that and it's the format has seen that I've got a cue Cal to format so you want that format to match the image that's come down you can mark it public or not this public flag here says can everybody in the OpenStack cloud use it or can only I use it in my project so that's that's a public private based on the the tenant or the project that is being imported in and then protected is a flag that says it can't be deleted unless that protected flag is taken off so the user that imported it can take that flag off or administrator can take it off but no one can delete it until that flag is is unchecked I'm going to do public just for fun create image now create image is not actually creating the image like the file itself all it's doing is creating a record of the file that I've previously downloaded so you download the file from the internet and then create image imports it into the registry I don't know why it didn't import okay quick check where you guys that it has anyone got a dashboard up and running okay good a few other rest you're still trying to get the vacant box oh how many you still try to get the vagin box up and running okay one of the challenges of a hotel Wi-Fi okay so for whatever reason it doesn't want to import photo right here and there's error messages we could go look at but we don't have time we've got zeros in there and so hopefully we'll be able to move forth and launch the Cirrus image okay so that was the process to import an image if you had a different image you wanted to import you could do that you can have multiple images in there you can even go to the extent of rolling your own image so if you wanted to build say a CentOS image that had your application pre-built into it so that when you launch an instance you can maybe launch over and over application inside of that image that you want to cluster together or you're doing maybe all your developers need a base image to do their development on you can custom roll these images and then add them to it and then your developers could come in and say yeah launch me another development environment launch me a different development environment and so managing these images can become very powerful for productivity in how you roll them and how you manage them again there's not time to be able to do that but there's lots of information online about how to create them so if you google you know create OpenStack cloud image or something like that then there's more information there yeah there's a line you can add I think did you get it didn't work in the back okay do you know what your solution was you have to show that there's nowhere to go so it's it's in here I was in the center fate the pagan box okay in the oops if you should have a bun - they're just mine's liver so it's probably in there this vagrant file here that's awesome to VirtualBox okay so so when vagrant pulls down that sent image we have a vagrant file that I've written that you're using to do this but there's also a vagrant file that describes the image the box that's been downloaded and so if you find where that box file is being stored on your machine there's a vagrant file next to it which is what I'm showing here and so you just need to change this our sink here to VirtualBox is that right VirtualBox and probably because I'm using a live bird provider but for VirtualBox provider you probably need to change that de VirtualBox you're going to need to find where your box file is so let's see so see how my box image file that's the CentOS box that it downloaded and that we've launched off of so this directory this home directory that I'm in may be different for a Windows machine but you need to find where that box image file is and where that vagrant file that came with it is and changed our stink to VirtualBox sorry I don't have better instructions on this okay so let me keep moving let's see we've created a user we've got an image imported into OpenStack that we can launch off of next we're going to create a network so once we have an image on a network then we can actually launch a VM so Neutron is the network management service and it's it creates virtual networks so we think of a network as a switch with a bunch of wires plugged into it and that creates the network you can do that virtually on servers and even have them span servers and this is what Neutron does it uses something called open V switch on the system and it ties all the open V switch services oarsman's services on the different machines together and then creates virtual networks so the same way that we think about a bunch of wires plugged into a switch physically open V switch can do that virtually with virtual machines and it can even segment them so that they're separate from one another the idea being here that will create a network and it goes into the project and that means that your project will have its own network that all the VMS can be attached to and no other project can attach to can attach to that Network only your VMs can unless you do extra configuration a quick question how many of you guys here use use VMware okay so so so the concept should be easy by so the open V switch is basically an open source version a similar to the theseare the vSphere distributor switch in a in a VMware environment right the virtual machines have virtual necks like on for the most part 2the typically don't can't plug into a direct link to a physical switch so you need some kind of a virtual switch that's really like a bridge that you can connect those virtual NICs to so that's all we're really doing so creating the network hop back in here select your network tab and select the network's link on the right I've got up in the top corner over here I've got to create networks create Network button I'm going to call this my private network and then you need to give it a subnet so this is kind of your your private networking subnet ranges I'm just going to do a 10 10 10 0 24 if you've used 1 9 2 6 8 or 1 72 16 those are those are good ranges to use and your you know we could just as easily put a 192 168 1 dot 0 slash 24 in there you don't need to fill out let's see I put that in the subnet name it should go in network address you don't have to put a subnet name you can if you want to you could name it private subnet there's there's a network and there's a subnet that goes with it so those are two different objects usually what I do is I name my network private and then don't name my subnet and just give it the the address that I'm using the gateway IP will be assigned automatically and a private network like this and then it's important to have DHCP enabled and that's by default on your network the reason is when the VM comes up the first thing it needs to do is get an IP address and so Neutron will statically assign an IP address to it but the instance will get it over DHCP ok just be sure just make sure everything's clear so what's happening is a every project slash tenant so I'm I'll probably use the word tenant just because it makes more logical sense you're old-school yeah but I really mean project so each project or tenant has to have its own network so think of again your from the VM world it's kind of like how you used to do V cloud 3 cloud director networking right so every tenant has its own kind of private network and for those VMs inside those networks they can only talk to each other to be able to talk to the outside world there's got to be a point like a provider network that's that tied into an you know a gateway to actually talks to the outside world so what will eventually do is basically connect a tenant network to one of those provider networks and that would allow the VMS actually talk outside its own little world and a subnet is just a range within a network a private network a range of IP addresses that a virtual machine can have so that make sense some way and you awake you need some jump in jet is it jumping that is a slide that said let's do jumping if you guys have questions because something's there's a lot information I know we're throwing at you you have questions just kind of raise your hand so we can kind of make sure that we're all on you know we're kind of all on the same page so just buy my book later yeah we could do that ii know it goes into your project so when I logged in as my user I'm logged into my project so all the virtual resources that I create what what the dashboard kind of hides for you is that when you log in you don't just log in as your user you log in as your user to a specific project and so I may be a user that's in multiple projects but when I'm authentic ating I'm authentic ating to a specific project so because I'm specific indicator that specific project all the virtual resources that I create automatically go into that project this so there's some division of responsibilities so as a Clow operator right you you're going to set up the underlying infrastructure because all this got to run on physical stuff right it may be a virtual network of virtual switches but it's got to sit on these guy actually got to talk to a real physical networking so you set that up you set up the provider network right that allows all the tenants would actually talk to the outside world but that each tenant has the ability to create his own private network and configure that and then basically say hey I just want to tie into this external network so I can talk outside so let's create an instance and put it on this private network and then we'll do the provider network side to show context ternal access so jump away from Neutron for a minute to Nova this is instance management this is basically the hypervisor manager that it's it's going to manage these virtual machines on demand across the hypervisors and OpenStack in general is intended to be built on standard hardware and beat it's designed to scale horizontally so we've put I've put this here that it's designed to scale horizontally and designed for standard hardware but that's OpenStack across the board not just Nova the intent is for you to go and take a bank of commodity servers stick OpenStack on it and really the only prerequisite that you have to have on these servers is that you have virtualization capabilities which pretty much all machines have now and then enough resources Ramen and CPUs to be able to divvy up into virtualization so let's take that network and the image that we have created that we imported and create an instance so I'm going to go compute into my instances and launch instance and we'll call it first instance because that's really creative and then your launch instance has all these tabs down the side so we need the ones that have the the blue stars or things that we have to go in and it has requirements for us so our source we have to go in and select this cirrus image and say this is the image we want to boot off of your flavor is a definition of how many how much resources are allocated to your virtual machine so you see the the pre-loaded ones there's tiny and small and medium and large and they have a certain number of V CPUs and RAM and disk that get allotted to them for this demo environment that we're basically doing nested virtualization because we're about to launch a VM inside of our VirtualBox VM or our vagrant VM just do tiny because if you do anything bigger it's not going to fit but I'm sorry how many of you use them AWS okay so this concept should be pretty familiar right basically doing very similar thing that you would do on an AWS environment so then next is networks I'm going to select the private network that we created and then there's a couple other tabs that aren't required right off the bat we'll jump back into a couple of them a little bit later here so now I'm going to do launch and my first instance comes up it's connecting to the hypervisor which this is kind of an all-in-one so it connects back to itself and it launches it builds and it should come up active so now you see this this active status that means that the virtual machine has gotten the disk image it's created a port on the network at spawn the VM it's come up and OpenStack sees it as a happy virtual machine ready for us to start interacting with right and the key is again keep in mind this is Dan is logged in as a regular user as a assumably a consumer of the cloud this is not something that work wise and operated a admin to be able to do right so the whole lot again the whole idea of OpenStack is giving end-users the power to create spin up their own resources just the way you would be able to do on it on Amazon Web Services but do it not only in a public cloud but potentially in a private cloud context so now we get back to that provider network idea that this instance has come up and it's on this private network but on that little private network in your project the only thing that that instance could talk to was if we spun up another instance then you could talk to that instance and it would be literally a little switch with two computers connected to it and they could talk to each other there's no internet access per se that's been provided to them so this provider network ends up being a a catch-all shared network that all tenants can end up connecting to and interacting with so we'll jump back into Neutron and look at creating a second network and a router to go with it and the router connects your private network to your public network so because this provider network is only available I'm sorry because the private net the public network is available for all projects to use we have to be an administrator to create an external network and I have that in quotes because that's kind of the flag that OpenStack gives these provider networks that when someone talks about an external network they're talking about one of these provider networks and you'll see in a minute when we create it there's a flag that says external and that's kind of where it came from so I'm going to log out from my non privileged user and log back in as the administrative user let me get my password back out of my Keystone RC file and you can change that password if you don't want to have to copy and paste it every time so now logging back in we can tell where the administrator again because we have this this admin panel here and this is where we're going to be able to manage this provider network scroll down and find networks and you'll see the private network that I created that says it's in my project is already listed there so now we're going to create a public network that we can then create a router to attach the public and the private network to one another so as administrator I'm going to create a network and I'm going to save public I like to put this in the services tenant because you're not supposed to attach a instance directly to the public network you need to have a router to go between and the services tenant or services project is a generic project that all the components get added to so that they're you know like I said every everything in OpenStack has to be inside of a project and the components are no different the public network is no different so this services project is kind of a catch-all don't really use this but it has to be in a project sort of designation and then down here at the bottom you'll see external network this is the designation that says this is a provider network so it's important that we check that network now the that wasn't supposed to happen hmm let's try Oh change your provider type to VX LAN and just put one as a segmentation ID external network I skipped a step read the manual yes VX LAN so by default neutron is configured to use VX land tunnels so that if you had multiple nodes it would connect those with these tunnels and then all of your tenant traffic would go across these tunnels we only have one node here it's an all-in-one so that's not terribly important but because the configuration by default is that it uses VX LAN we have to specify that our public network is of type VX lambda so one of the powerful things about OpenStack is that there are so many options there's so many ways I mean that you know the fact that you can pick all these different network types from these virtual network types to actually just regular straight-up VLANs right in flat network that's a very powerful thing it's also one of the most painful things about OpenStack is that you know there's there's 300 different dead type 300 thousand different combinations ways you can configure OpenStack and about a couple hundred of them actually working production so that's that's one of the challenges and one of the reasons why there are people offering OpenStack as as a distribution or service because they've they're basically kind of giving their opinion Aidid their opinion on what is the right configuration that could actually work in the production environment but for the purpose of playing with it it's good to try to play as many difficulties as many of these different options as possible to see what you can actually do and what actually works and everybody starts in the same place so just because it feels like you're drinking from three fire hoses worth of information when you first start trying to configure OpenStack cloud know that everybody started I started there Ken started there everybody has to go through kind of the tough process of learning enough configuration options to get far enough ahead to be able to bring something up and use it and so hopefully this vagrant file can help you at least get started with that and be able to interact with it so that you have a baseline to move from now notice is the administrator that I created the public network in the services tenant but there's no subnets associated with it the automatic subnet creation is a non privileged user feature within the dashboard as an administrator you have to create them separately from one another so I'm going to select this public network and it gives me the option to go in and create subnets so you see over on the right side now I'm going to click this create subnet button and here again the subnet name you can name it if you want I generally just name the networks and not the subnets you're welcome to do either the network address that you want to use is this network address let me get it typed in and then I'll switch back so that you can that you can copy it off and the way that the vagrant file is designed we have to you have to use this one specific to the vagrant file because that's the way I configured it but know that provider networks are something that are generally provided by your network administrator so a lot of times this information would be given to you from a network administrator and they would say use the specific Sider because this is the block of IP addresses that I've given you for your OpenStack cloud to work it does okay let me finish this real quick and then I'll go back and give you a sec to look at that slide on the subnet we need to go into subnet details oh I put it in the name again maybe I should just start giving the subnets name so I put them at the right box okay the the important thing about creating an external network is that you should disable DHCP so by default your internal private networks are going to have DHCP so that when your instances come up they get DHCP off of your network that you've created but because a provided or network is one that your network administrator has given you this information you're using a network that's been provided to you from your network administrator so you don't want to put DHCP on that network because there's probably a DHCP service already running and you don't want those to conflict and the IPS that you've been given by your network administrator will be assigned statically so be sure to enable a disable DHCP on it and then oftentimes there's a a subset of these IPS that need to be used and that's called an allocation pool so here at the bottom I also have an allocation pool and all that says is the IPS that this provider network will allow you to use are in the range from this you know four to twenty seven to four to thirty eight so it's it's saying even though you have a slash twenty eight only use a certain amount of the IP addresses so I've got to tighten out okay does anyone still need this up for a few minutes everybody good for now great I have a link at the end and actually the link that goes to the vagrant file if you just take vagrant file off the back of it and my fedora people drive there you can look for the Austin PDF and there's a PDF out there that has all the slides on it right next to the in the same place as the vagrant file yeah same exact slides that are here in that PDF it was out your question was I have information yes that slide is in there yeah so everything that's going up here is in that PDF so that slide will be in there with that information it's hard to see with the light okay so now I've created the provider network now we want to attach the private network that our instance is attached to out to that provider network so that the instance has external connectivity so I'm going to log back out from the administrator account and back into my non-privileged account and there's kind of a neat thing that they have this network topology link under networks will bring up kind of a visual representation of what the network looks like so this guy right here is the provider network that we just created and this little cloud here is the the pub the Private Network did I say that right the provider network is the little globe right and the private network is the little cloud and then the instance is the little Mac it looks like a Mac doesn't it in your non privileged user account there's a network tab so drop down the network and there's network topology so what we need to connect this provider networks to the private network is a router so if we go down here to routers if I only have one router I generally will name it the same thing as my project and then it gives you a external network option here so I'm going to go ahead and attach the public network to this router and then you'll see once it's created over on the right side it says clear gateway so another name for this provider network being attached to your router is a gateway so if I were to hit clear gateway it would detach those two and it would say set gateway and I would be able to select that public network again so now if we look at the the network topology again we see that the the globe the provider network is connected to a router these arrows but we still don't have a connection back into the private network to create a route all the way out so go back to your router and select the router and the connection from the private interface the private network into the router is called an interface so if I select the interfaces tab on that router say add interface and select my private network from the list and submit it now I have a connection from the private network to the router and the public network to the router and we should be able to see that in our visualization here so you can see there's a link from the instance to the private network to the router to the public network and this public network is kind of in quotes public right that if you get actual public IPs on this public network then there's actual public access like real world internet access but public is a little bit misrepresented there in that if you have a corporate network you still need this public network that has IPS on your corporate network for you to get into your internal network so internal means that it's isolated from everything and public is or the provider network the external network is the external connection outside of the OpenStack cloud to whatever network it's connected to so this could be like I said a corporate network that you're on it could be your home router you know if you're doing this at your house and you need to get in through your 192 168 home router addresses it could be actually public IPS that are that are you know legit IPS that you can get from anywhere in the world so at this point we configured that so at this point you should be able to SSH into that instance I'm sorry there's a step before that here iñaki yeah we need to assign a floating IP so a floating IP is an IP from that provider network that's assigned to the instance so that you can then contact it so right now my Cirrus instance has a a 10 10 10 3 address and if I come over here and say associate floating IP then I can add one to that instance now a floating IP is a resource just like everything else so I can't just say give me a floating IP to my instance I have to first allocate a floating IP into my project and then take the floating IP that's been allocated to the project and associate it to the instance so here it says manage floating IP associations with this instance but it says no floating IP addresses alakay so the dashboard is great about this it gives you a little plus button it says let's allocate a floating IP it'll come from the public network associate IP now there's an IP that is listed in my IP address list there and I can say actually associate that with first instance the name of the instance that we've launched over on the right side of the screen there's a button next to create snapshot and this disassociate floating IP was associate floating IP so each of the each of your instances that come up will have a drop-down that have options of different sources attic so these are the IP addresses remember back when you were creating the provided network you put in the cider and there was a IP address range all we're doing here is say hey I I know you have this range of IP addresses that was created it was signed give me one of those so that I can give it to one of my instances so then your instance typically will have two virtual necks right one neck will have that private network IP address so you can only talk internally and then the other neck you're going to assign this floating that's what you're going to connect to the external or providing network that one gets the floating IP address and it's what the iraq should be allow you to talk to the outside talk outside your tenant into two Ida to other tenants or even to the external world so someone had a question the link that you got the vagrant file from it's in that same directory so if you do radius Fedor people Oregon just get the directory listing of that there's a PDF in that directory there right next to the vaguer file cool okay so now that we have that floating IP associated with the instance I'm on my laptop and the networking vagrant has set up the networking for us to get from my laptop into this virtual machine so you can see the top the 4.2 28 that was associated with the instance I'm able to ping it and then at the bottom here I was able to SSH into my Sierra's image by default cirrus wants you to login as user zeros one of the security things they tried to get right and then you'll notice down at the bottom though that when I log in it's asking for a password in general when you download a VM or a cloud image from the internet they're not going to give you a username and password to get into it so you have to setup a key pair an SSH key pair so in the compute tab under access and security you can manage your SSH key pairs and you can create a key pair similar to the way you do in AWS where it'll create one and it will force you to use a private key that you download OpenStack also allows you to import a key pair so I'm actually just going to import my public key that's on this laptop it was it's a virtual router which is handled by network namespaces at the Lenox level so there's a there's a DHCP network namespace and there's a router Q router deep namespace and so if you if you search for those if you want to know more about them that's some of the underlying plumbing of network namespaces and OBS that's being done under the covers okay so I've I've imported this key pair now I'm not sure that I'm able to run two instances on this demo environment once so what I'm going to do is terminate my first one so I'm going to go into that same drop down and confirm delete instance because when this one booted it didn't add the the key pair wasn't associated that was one of the options in launch instance so I'm going to go back into my instance launch and I'll do second instance again cuz I'm really creative and apparently not very funny either so couple of me while he's doing this this pot how many you feel like this is that overly complicated it seems these other things can very overly complicated just set things up no one that's good for user yeah so a couple of things to keep in mind what is that Dan is doing a lot of the kind of the pre prep work right to get things up up and running so in a steady state operational environment you shouldn't have to keep doing this over and over that's the one thing second thing is in my experience is if you're if you're users or developers particularly almost none of them will ever use this dashboard but using this dashboard because because an easy way to visualize how to do things in a workshop but in reality your developers are most likely be using command line and POW and if they hesitate you say we're doing it right they're going to be using the API and using different tools to actually provision these resources so you're not they're not going to be clicking through a bunch of tabs and wizards everything should be laid out in a config' you know some kind of a manifests or the config file or something in the way they code their applications and then the third thing is if you haven't picked this up already to see if it seems like allow you folks are having a vmware background which i also have is you to learn Linux there's no way around it it's especially on the networking piece if you don't understand namespaces you can have you know those things like an IP tables I you're going to struggle that's the only way I can put it so you know I'm not saying you have to be in Linux school but you have to low enough about Linux concept particularly the networking space to be able to afford a lot of stuff to make sense for you as you're operating it okay so here launching this instance I selected all the same source flavor networks and everything the difference is I went to this key pair tab and I'm going to select my key pair that I just imported and then launch the instance and all that's going to do is when the instance launches it connects into a metadata service and pulls that key pair out and drops it in as the user so once this guy comes up I'd be able to log in without having to put in that password I'd just be able to SSH directly to it we're running out of time here so I'm going to go ahead and move and not actually display that we've got two things left real quick to talk about cinder block storage again I talked about earlier if you if your instance gets ax for some reason you want to have a place that you can save in save off data so that if it does get axed you can reattach it to another VM and access that data so cinder does this for us there's a volumes link that I just clicked here over on the right again we can create volume give it a name like first of all the source and the type aren't necessary they're kind of extra parameters the size is what you want to pay attention to so I'm just going to create a one gig volume but if you had a larger backing then you could do 10 or 20 100 Gig whatever you're you're able to provide and then once that volume is created you simply attach it so manage attachments over here on the right and it'll give you a list of your instances so I'm going to select my instance that I have running and say attach volume and so now if you logged into that VM the initial drive was dev SDA or VD a I think for for virtual disc a and then when this guy gets attached which it says it's attached now another one will pop up inside of there and it will say V DB so you now have a second block device but you can now create a file system in a partition table and mount into the Linux file system just as you would had you plugged like a physical drive into a desktop or a server of some sort yeah so because they're going to okay you want to skip Swift just because the time okay because you asked ticking ask questions right so because we're willing suicides just keep it keep in mind by default when you create a VM and the the root disk it's an ephemeral disk that means when you just when you delete the instance all the data gets purged right and you can imagine in some cases that's not what you want that's where the cinder block storage project comes in that's a way for you to take a a volume attach it like as Dan set almost like a USB Drive and then whatnot in case when you terminate the instance the data the values persists over past that time and still has all the data on it and you can read it and you can attach it to something else one thing I need to caution you about is although a lot of customers use that kind of sentence and like that traditional emc or NetApp array to craft to provide the sender the backend for this in the volume cinder is not a share storage technology right this is not like vmware where you can take an e you know of value of value from EMC and say ok now I get to VM seconds that can both connecting the cluster and then 1 1 hypervisor fails we started the datas not the way works when if the hypervisor that you attach the ascender volume to dies it you have to manually your script detaching that volume and reattach it to something else okay that's really important is that I always get people get confused because they in the head they think it's like a VMware type thing but there's no cluster throughout the project Manila which does share a style system so I know the senders like a sand and Manila is shared file systems like NFS or something like that I just labeled myself as an Linux guy NFS it's like okay so the final thing so we're gonna just for time sake we're going to skip Swift you can read about object storage if you search online you know using it in the dashboard is is pretty simple just like everything else there's a create container add object in container there's those are the two concepts is there's a container that you add files to and then you add objects which are just simple file there's no metadata about these files there's just a name and the content that goes with it I want to touch real quick from a Red Hat perspective I mentioned earlier that we have kind of the community and the the supported product separation within the RTO community we have two installation methodologies one is called pack stack which is what we use today and it's generally intended for demonstration proof of concept small deployments where you're kind of playing and learning OpenStack the other one that we have out there is called triple o and there's a QuickStart out there that you can read through and work on that as well if you'd like to get involved or to try and use triple o triple O's basis is that triple oooo stands for OpenStack on OpenStack and the idea is that it uses OpenStack to deploy OpenStack and so what it actually does it stands up an all-in-one OpenStack just like the one we're using today here but it adds in bare metal support which is the project ironic and so the bare metal support can then go out and provision a larger cluster of machines and triple o is where you're going to get all of your your bigger capabilities like high availability or like provisioning larger software-defined storage clusters and things like that and then from the enterprise side the supported side OSP director is our supported product and so the thing to note here is that triple o and OSP director are one one to one with each other triple o is our Community Supported installer and OSP director is our our productized supported product so if you pick up triple o and you like what's in triple o or need help with triple o and some rate in some aspect oh s P director is what we're going to be is what we sell as a product and so it's important to know that PAC stack is kind of just for play and triple o is is intended for a larger longer-term supported type installation this is just a review of the first slide we have with all the different components that we've worked through and then here's a resources page with you know visit Rackspace Rackspace and Red Hat are partnering now to provide cloud solutions are do project is where you can get the are do bits and the community support of what we've worked on here today OpenStack of course try stack is a free platform that you can go and spawn instances as a non privileged user so interacting with the networking and the images and the block storage and the object storage if you don't want to have to install OpenStack but you want to use a demonstration type environment try stack as a set of servers funded by the foundation and managed by Red Hat so exactly what we just used here in this demonstration is what's running on a bank of servers at try stack org and then my fedora people link at the bottom is where the vagrant file on the PDF are so you're welcome to pull those down and you know email if you're having trouble with those I think that's it for us if you have questions you're welcome to stick around and ask them I do have time for a couple not really right Tyler yeah if you have questions come on up and see us and we'll stick around for a little bit thanks for coming guys thanks for hanging with us for nice long session

Channel: Open Infrastructure Foundation

Views: 58,694

Keywords: OpenStack, OpenInfra, Open Infrastrucure, Open Source, Demo

Id: -xsvYo0_cZg

Length: 87min 29sec (5249 seconds)

Published: Sat Apr 30 2016